Jensen Huang, founder and CEO of Nvidia, recently highlighted his company’s AI advancements in his keynote at COMPUTEX 2024, stressing the importance of generative AI and its ability to create new content.
In this article, I would like to explore how Nvidia is using these technologies to revolutionize the understanding of the world, enabling with its technology the creation of detailed digital clones and replicas, which will undoubtedly reshape industries, products, and the way we study the world around us, and all in the form of tokens.
I was honestly blown away by Nvidia’s breakthrough in the area of infrastructure that drives the progress of Artificial Intelligence. The two technologies that drive these developments are the exponential acceleration of computing where precisely this manufacturer is playing a predominant role and generative artificial intelligence. generative artificial intelligence.
The launch of its latest super-chip, Blackwell, presented at this event, has left Moore’s law at the height of the bitumen, as you can see in this graph. This is called exponential acceleration . This super-chip has also reduced the power consumption for token generation also exponentially ensuring the scalability of the processing of generative AI models.
We are facing "The Era of Generative Artificial Intelligence".
I am convinced that you will find the new “Omniverse” they have created exciting for the progress it represents in the field of real-world simulation.
Let us now take a look at the new products presented by the company
Nvidia's NIMS
NVIDIA Inference Microservices (NIMs) are pre-built, enterprise-ready containers that include NVIDIA inference software, such as Triton Inference Server and TensorRT-LLM, to accelerate the deployment of AI models in production.
These micro-services enable developers to build generative AI applications for copilots, chatbots and digital avatars, and can be integrated into platforms such as Hugging Face and Amazon SageMaker.
They can be connected to each other, to create expert teams, just as a team of humans would.
Advantages
- Speed and efficiency: enable developers to build generative AI applications in minutes instead of weeks.
- Flexibility: Pre-built containers allow developers to scale and deploy AI applications across different infrastructures, such as clouds, data centers, and workstations.
- Access to AI models: offer access to more than 40 pre-trained AI models, including Meta Llama 3, Microsoft Phi-3 and Mistral Large.
- Copilots and chatbots: building copilots and chatbots that can interact naturally with users
- Digital avatars: construction of digital avatars that can interact with users in a realistic way.
- Smart factories: enable factories to automate and improve their efficiency by integrating autonomous robots and machine vision.
Digital Humans
NVIDIA ACE is a suite of generative AI technologies that enables the creation of digital humans with realistic conversational interactions. These digital humans can interact with users in a natural way, using microservices such as speech recognition, language generation, and facial animation.
Companies such as Hippocratic AI, Inworld AI, and UneeQ are using ACE to develop virtual assistants, game NPCs, and customer service avatars.
World Simulations
NVIDIA Omniverse is a 3D simulation and authoring platform that allows you to integrate data from multiple CAD applications, and create photorealistic digital twins. These simulations allow manufacturers to virtually optimize their robotic factories, reduce costs, and improve efficiency.
Electronics manufacturers such as Foxconn, Delta, Pegatron and Wistron are using Omniverse to virtually design and optimize their robotic factories. They can simulate the layout of equipment, workflows, autonomous robots, and AI vision systems before implementing them in reality.
Can you imagine being able to simulate, for example, the planet Earth itself using vast amounts of data from a variety of sources, including satellites, weather stations and ground sensors to create a “digital twin” of the real world, used for everything from urban planning to natural disaster mitigation?
Smart Factories
With its advanced AI and simulation technologies, Nvidia enables the creation of digital factory twins. These twins help optimize the design and operation of facilities before their physical implementation by simulating their operation, allowing decisions to be made before they are implemented in the real world.
NVIDIA’s Isaac platform facilitates the development and deployment of autonomous robots, while Metropolis focuses on industrial machine vision for applications such as automated optical inspection. These innovations allow manufacturers to improve operational efficiency and reduce costs without making physical changes to the facility.
For example BMW has launched the first virtual factory in the Nvidia Omniverse. Do not miss the following video about it.
Impact and Practical Applications of these technologies
NVIDIA revolutionizes the way we interact with technology, making it more efficient, secure, and accessible.
Developers can reduce deployment times from weeks to minutes, create generative AI models that interact naturally with users, and automate processes in smart factories, allowing humans to focus on higher, more creative tasks.
It also allows the simulation of environments and processes in a virtual environment, which facilitates the planning, design and optimization of complex systems, reducing costs and risks, and improving efficiency and safety in various industries.
I honestly believe that the company has undoubtedly been the engine that has brought us to the point where we are today in the field of generative AI. We will soon see these solutions, currently available in data centers, extend in some form to our portable devices, giving us access to the enormous power of AI on them.
Interesting, isn’t it?
And what do you think of these developments, I’d love to hear from you in the comments 🙂
Have a good week!
References of interest:
- NVIDIA Digital Human Technologies Bring AI Characters to Life
- NVIDIA Releases Digital Human Microservices, Paving Way for Future of Generative AI Avatars
- Robotic Factories Supercharge Industrial Digitalization as Electronic Makers Adopt NVIDIA AI and Omniverse
- NVIDIA Omniverse
- NVIDIA NIM Revolutionizes Model Deployment, Now Available to Transform World’s Millions of Developers Into Generative AI Developers
- NVIDIA to Build Earth-2 Supercomputer to See Our Future
